Can the minimal clinically important difference be determined in a French-speaking population with primary hip replacement using one PROM item and the Anchor strategy?

BACKGROUND

The impact of surgery on the patient is classically assessed on pre- and post-treatment scores. However, it is increasingly recommended to rank these results according to the minimal clinically important difference (MCID), using either the data distribution method or the anchor method, latter consisting in an extra question specifically targeting the patient's improvement. MCIDs vary between populations and, to the best of our knowledge; there have been no investigations in France regarding this in the context of total hip replacement (THR). Therefore, we conducted a prospective study in a population with THR to determine: 1) whether MCID scores in France were comparable to those reported in the data from the international literature; 2) whether a general item taken from a different score could serve as an anchor; and 3) whether an item from the actual questionnaire itself could serve as an anchor.

HYPOTHESIS

When pre- and post-treatment scores are available, an item from the questionnaire itself can serve as an anchor for MCID.

MATERIAL AND METHODS

In a prospective observational study, 123 primary THR patients (69 male, 54 female), out of 150 initially included, completed the 5 domains of the HOOS hip disability and osteoarthritis outcome score and the Oxford-12 questionnaire, preoperatively and at 6-12 months. The MCID was calculated via the distribution-based and the anchor-based methods. Two Oxford items (questions 1 and 2) and 2 HOOS items (questions S1 and Q4) were used as anchors, as well as a supplementary question on improvement and the Forgotten Joint Score (FJS).

RESULTS

At a mean 10.12±1.2 months' follow-up [range, 6.5-11.9 months], the Oxford-12 score increased from 19±8 [3-35] to 40±10 [8-48] (p<0.001), all HOOS components demonstrated improvement, and the FJS at the final follow-up was 71±29 [0-100]. The general items (Oxford question 1 and HOOS question Q4) were more discriminating than the joint-specific items (Oxford question 2 and HOOS question S1). Based on results from the 3 anchors (improvement rated 1 to 5, Oxford question 1 and HOOS question Q4), 3 to 5 patients showed deterioration, 5 to 6 were unchanged, 30 to 40 were slightly improved, and 73 to 80 were improved by THR. The mean MCID on both distribution and anchor methods was 9 [5.5-12] for Oxford-12, 20 [12-27] for HOOS symptoms, 26 [10-36] for HOOS pain, 22 [11.5-28] for HOOS function, 26 [13-34] for HOOS sport and 22 [14-28] for HOOS quality of life.

DISCUSSION

The MCID for the Oxford-12 and HOOS scores in a French population was comparable to data from the past literature. Using a score item as an anchor to define improvement is possible, but only if a general item is used.

LEVEL OF EVIDENCE

IV; prospective study without control group.

CLINICAL TRIALS REGISTRATION

NCT04057651.